Using Supervised Learning to Select Intraday Volatility Models
Summary
The document introduces the challenge of forecasting intraday volatility across financial assets. It notes that researchers and practitioners have developed many approaches, including intraday estimators, GARCH models, and continuous-time models. Volatility forecasts can inform trading signals, algorithm design, and portfolio allocation, so model choice can affect several parts of a quantitative process.
Its proposed direction is to use supervised learning to automate the selection of a suitable volatility model for each asset, with monitored indicators used to revise the choice as conditions change. The discussion argues that a single model may not work well across different asset types and that manual selection can be complex and inefficient. However, the available text is only an overview: it supplies no model specifications, training data, evaluation results, or details about the monitoring and correction process. The claims about accuracy therefore cannot be assessed from this document alone.
Key ideas
- Volatility forecasts can support signals, algorithmic strategies, and portfolio allocation.
- Different asset types may require different volatility models.
- Supervised learning is proposed as a way to automate model selection.
- Monitoring indicators could help revise model choices as market conditions change.
- The document gives no empirical results or implementation details.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.